<![CDATA[Purpose – This study aims to develop and evaluate an IoT-enabled, rule-based adaptive traffic signal prototype that responds to discrete vehicle queue states at a four-way intersection. The system is designed as a low-complexity alternative to fixed-time signal control by associating detected vehicle presence at predefined spatial points with different green-light durations.Methods – The prototype was developed using ESP32 microcontrollers and eight infrared (IR) sensors, with two detection points assigned to each of four approaches. The physical prototype uses spatially separated detection points to represent different queue states. Sensor inputs are processed locally by the ESP32 and classified into predefined traffic states. The controller assigns green-light durations of 20, 25, or 30 s according to the detected state. The prototype was evaluated under five laboratory-scale traffic configurations using simulated vehicles to verify sensor detection, traffic-state classification, signal response, and IoT communication functionality.Findings – The experimental evaluation verified that the controller responded according to the predefined queue-state rules. When neither detection point was active, the corresponding traffic state was classified as normal and a 20-s green duration was assigned. Activation of the first detection point resulted in a 25-s green duration, while activation of both detection points resulted in a 30-s green duration. The five predefined scenarios demonstrated that the controller could apply different timing levels according to the traffic-state distribution among the four approaches. Repeated-cycle testing resulted in 49 successful detection-and-control responses out of 50 cycles, corresponding to a 98% successful-cycle rate under the tested laboratory conditions.Research Implications – The results demonstrate the functional feasibility of integrating discrete IR-based queue-state detection, local ESP32 processing, traffic-signal control, and wireless communication in a laboratory-scale prototype. However, the study does not establish a quantitative reduction in real-world delay, queue length, or congestion because the evaluation does not include field testing or a comprehensive fixed-time performance baseline.Originality – The contribution of this study lies in the implementation of a low-complexity two-point queue-state detection architecture that translates spatial vehicle-presence information into rule-based green-time selection on a four-approach prototype. The architecture provides a basis for subsequent development involving repeated quantitative experiments, additional detection zones, improved traffic-state estimation, and field-based evaluation.]]>
Copyrights © 2026
